virtio-pci registers a per-vq affinity hint when using MSIX,
but fails to remove it when freeing the interrupt, resulting
in this type of splat:

[   31.111202] WARNING: CPU: 0 PID: 2823 at kernel/irq/manage.c:1503 __free_irq+0x2c4/0x2c8
[   31.114689] Modules linked in:
[   31.116101] CPU: 0 PID: 2823 Comm: kexec Not tainted 4.10.0+ #6941
[   31.118911] Hardware name: Generic DT based system
[   31.121319] [<c022fb78>] (unwind_backtrace) from [<c0229d8c>] (show_stack+0x18/0x1c)
[   31.125017] [<c0229d8c>] (show_stack) from [<c05192f4>] (dump_stack+0x84/0x98)
[   31.128427] [<c05192f4>] (dump_stack) from [<c023d940>] (__warn+0xf4/0x10c)
[   31.131910] [<c023d940>] (__warn) from [<c023da20>] (warn_slowpath_null+0x28/0x30)
[   31.135543] [<c023da20>] (warn_slowpath_null) from [<c0290238>] (__free_irq+0x2c4/0x2c8)
[   31.139355] [<c0290238>] (__free_irq) from [<c02902d0>] (free_irq+0x44/0x78)
[   31.142909] [<c02902d0>] (free_irq) from [<c059d3a8>] (vp_del_vqs+0x68/0x1c0)
[   31.146299] [<c059d3a8>] (vp_del_vqs) from [<c056ca4c>] (pci_device_shutdown+0x3c/0x78)

The obvious fix is to drop the affinity hint before freeing the
interrupt.
Signed-off-by: NMarc Zyngier <marc.zyngier@arm.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This reverts commit 5c34d002.

Conflicts:
	drivers/virtio/virtio_pci_common.c

The cleanup seems to be one of the changes that broke
hybernation for some users. We are still not sure why
but revert helps.

This reverts the cleanup changes but keeps the affinity support.
Tested-by: NMike Galbraith <efault@gmx.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This reverts commit 07ec5148.

Conflicts:
	drivers/virtio/virtio_pci_common.c

Unfortunately the idea does not work with threadirqs
as more than 32 queues can then map to a single interrupts.

Further, the cleanup seems to be one of the changes that broke
hybernation for some users. We are still not sure why
but revert helps.

This reverts the cleanup changes but keeps the affinity support.
Tested-by: NMike Galbraith <efault@gmx.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This reverts commit 53a020c6.

The cleanup seems to be one of the changes that broke
hybernation for some users. We are still not sure why
but revert helps.
Tested-by: NMike Galbraith <efault@gmx.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This reverts commit 52a61516.

Conflicts:
	drivers/virtio/virtio_pci_common.c

The cleanup seems to be one of the changes that broke
hybernation for some users. We are still not sure why
but revert helps.

This reverts the cleanup changes but keeps the affinity support.
Tested-by: NMike Galbraith <efault@gmx.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This reverts commit de85ec8b.

Follow-up patches will revert 07ec5148 ("virtio_pci: use shared
interrupts for virtqueues") that triggered the problem so no need for
this one anymore.
Tested-by: NMike Galbraith <efault@gmx.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Fedora has received multiple reports of crashes when running
4.11 as a guest

https://bugzilla.redhat.com/show_bug.cgi?id=1430297
https://bugzilla.redhat.com/show_bug.cgi?id=1434462
https://bugzilla.kernel.org/show_bug.cgi?id=194911
https://bugzilla.redhat.com/show_bug.cgi?id=1433899

The crashes are not always consistent but they are generally
some flavor of oops or GPF in virtio related code. Multiple people
have done bisections (Thank you Thorsten Leemhuis and
Richard W.M. Jones) and found this commit to be at fault

07ec5148 is the first bad commit
commit 07ec5148
Author: Christoph Hellwig <hch@lst.de>
Date:   Sun Feb 5 18:15:19 2017 +0100

    virtio_pci: use shared interrupts for virtqueues

The issue seems to be an out of bounds access to the msix_names
array corrupting kernel memory.

Fixes: 07ec5148 ("virtio_pci: use shared interrupts for virtqueues")
Reported-by: NLaura Abbott <labbott@redhat.com>
Signed-off-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Tested-by: NRichard W.M. Jones <rjones@redhat.com>
Tested-by: NThorsten Leemhuis <linux@leemhuis.info>

This basically passed up the pci_irq_get_affinity information through
virtio through an optional get_vq_affinity method.  It is only implemented
by the PCI backend for now, and only when we use per-virtqueue IRQs.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Add a struct irq_affinity pointer to the find_vqs methods, which if set
is used to tell the PCI layer to create the MSI-X vectors for our I/O
virtqueues with the proper affinity from the start.  Compared to after
the fact affinity hints this gives us an instantly working setup and
allows to allocate the irq descritors node-local and avoid interconnect
traffic.  Last but not least this will allow blk-mq queues are created
based on the interrupt affinity for storage drivers.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Try to grab the MSI-X vectors early and fall back to the shared one
before doing lots of allocations.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This lets IRQ layer handle dispatching IRQs to separate handlers for the
case where we don't have per-VQ MSI-X vectors, and allows us to greatly
simplify the code based on the assumption that we always have interrupt
vector 0 (legacy INTx or config interrupt for MSI-X) available, and
any other interrupt is request/freed throught the VQ, even if the
actual interrupt line might be shared in some cases.

This allows removing a great deal of variables keeping track of the
interrupt state in struct virtio_pci_device, as we can now simply walk the
list of VQs and deal with per-VQ interrupt handlers there, and only treat
vector 0 special.

Additionally clean up the VQ allocation code to properly unwind on error
instead of having a single global cleanup label, which is error prone,
and in this case also leads to more code.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

We don't really need struct virtio_pci_vq_info, as most field in there
are redundant:

 - the vq backpointer is not strictly neede to start with
 - the entry in the vqs list is not needed - the generic virtqueue already
   has list, we only need to check if it has a callback to get the same
   semantics
 - we can use a simple array to look up the MSI-X vec if needed.
 - That simple array now also duoble serves to replace the per_vq_vectors
   flag
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

There is basically no shared logic between the INTx and MSI-X case in
vp_try_to_find_vqs, so split the function into two and clean them up
a little bit.

Also remove the fairly pointless vp_request_intx wrapper while we're at it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

vp_request_msix_vectors is only called by vp_try_to_find_vqs, which already
calls vp_free_vectors through vp_del_vqs in the failure case.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

This avoids the separate allocation for the msix_entries structures, and
instead allows us to use pci_irq_vector to find a given IRQ vector.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

We call del_vqs twice when request_irq fails, this
makes no sense.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Introduce PCI_VENDOR/PCI_SUBVENDOR/PCI_SUBDEVICE defines to replace the
constants scattered in the kernel already used to detect QEMU.

They are defined in the QEMU codebase per docs/specs/pci-ids.txt.
Signed-off-by: NRobin H. Johnson <robbat2@gentoo.org>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Reviewed-by: NGerd Hoffmann <kraxel@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Acked-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

KASan detected a use-after-free error in virtio-pci remove code. In
virtio_pci_remove(), vp_dev is still used after being freed in
unregister_virtio_device() (in virtio_pci_release_dev() more
precisely).

To fix, keep a reference until cleanup is done.

Fixes: 63bd62a0 ("virtio_pci: defer kfree until release callback")
Reported-by: NJerome Marchand <jmarchan@redhat.com>
Cc: stable@vger.kernel.org
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Tested-by: NJerome Marchand <jmarchan@redhat.com>

checkpatch.pl wants arrays of strings declared as follows:

  static const char * const names[] = { "vq-1", "vq-2", "vq-3" };

Currently the find_vqs() function takes a const char *names[] argument
so passing checkpatch.pl's const char * const names[] results in a
compiler error due to losing the second const.

This patch adjusts the find_vqs() prototype and updates all virtio
transports.  This makes it possible for virtio_balloon.c, virtio_input.c,
virtgpu_kms.c, and virtio_rpmsg_bus.c to use the checkpatch.pl-friendly
type.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Acked-by: NBjorn Andersson <bjorn.andersson@sonymobile.com>

Move resource allocation from common code to legacy and modern code.
Only request resources actually used, i.e. bar0 in legacy mode and
the bar(s) specified by capabilities in modern mode.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

The cpumask vp_dev->msix_affinity_masks[info->msix_vector] may contain
staled information when vp_set_vq_affinity() gets called, so clear it
before setting the new cpu bit mask.

Cc: stable@vger.kernel.org
Signed-off-by: NJiang Liu <jiang.liu@linux.intel.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

The PCI core now disables MSI and MSI-X for all devices during enumeration
regardless of CONFIG_PCI_MSI.  Remove device-specific code to disable
MSI/MSI-X.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

If set, try legacy interface first, modern one if that fails.  Useful to
work around device/driver bugs, and for compatibility testing.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

Lightly tested against qemu.

One thing *not* implemented here is separate mappings
for descriptor/avail/used rings. That's nice to have,
will be done later after we have core support.

This also exposes the PCI layout to userspace, and
adds macros for PCI layout offsets:

QEMU wants it, so why not?  Trust, but verify.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Most of initialization is device-independent.
Let's move it to common.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

It turns out we need to add device-specific code
in release callback. Move it to virtio_pci_legacy.c.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Our code calls del_vqs multiple times, assuming
it's idempotent.

commit 3ec7a77b
    virtio_pci: free up vq->priv
broke this assumption, by adding kfree there,
so multiple calls cause double free.

Fix it up.

Fixes: 3ec7a77bReported-by: NSasha Levin <sasha.levin@oracle.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

When the virtio_pci driver was moved into virtio_pci_legacy.c the module
licence and other attributes went AWOL.  This patch restores them.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

It turns out this make everything easier.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

kbuild does not seem to like it when we name source
files same as the module.
Let's rename virtio_pci -> virtio_pci_common,
and get rid of #include-ing c files.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

There's been a lot of changes since 2007.
List main authors, add Red Hat copyright.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Move everything dealing with legacy devices out to virtio_pci_legacy.c.
Expose common code APIs in virtio_pci.h
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

config vector setup is version specific, do it indirectly.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

VQ setup is mostly version-specific, add another level of indirection to
split the version-independent code out.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

VQ deletion is mostly version-specific, add another level of indirection
to split the version-independent code out.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

slightly reduce the amount of pointer chasing this needs to do.
More importantly, this will easily generalize to virtio 1.0.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

We don't need to go from vq to vq info on
data path, so using direct vq->priv pointer for that
seems like a waste.

Let's build an array of vq infos, then we can use vq->index
for that lookup.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

should be

struct foo {
}

not

struct foo
{
}
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>